One of the biggest problems in the correlation of FE models and data from experimental modal analysis (=MA) turned out to be the large difference in the dimension of the two models. In this paper several expansion methods that allow the expanding of the eigenvectors of the MA model to the dimension of the FE-model are compared with each other and with the "correct" solution by simulating the measurement in a second FE analysis. The expanded eigenvectors are then used in an error localization process (the modal force error method was applied) in order to test them on their capability to localize the true error spot(s). The test object is a simple spring mass structure. The results indicate that no expansion method is the best for all cases, but some trends are clearly observable.
展开▼
